OWLv2: bug fix in post_process_object_detection() when using cuda device (#27468)

* OWLv2: bug fix in post_process_object_detection() when using cuda device

* fix copies issue by fixing original function in owlvit

src/transformers/models/owlv2/image_processing_owlv2.py

@@ -504,7 +504,7 @@ class Owlv2ImageProcessor(BaseImageProcessor):
             else:
                 img_h, img_w = target_sizes.unbind(1)
 
-            scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
+            scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
             boxes = boxes * scale_fct[:, None, :]
 
         results = []

src/transformers/models/owlvit/image_processing_owlvit.py

@@ -448,7 +448,7 @@ class OwlViTImageProcessor(BaseImageProcessor):
 
         # Convert from relative [0, 1] to absolute [0, height] coordinates
         img_h, img_w = target_sizes.unbind(1)
-        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
+        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
         boxes = boxes * scale_fct[:, None, :]
 
         results = [{"scores": s, "labels": l, "boxes": b} for s, l, b in zip(scores, labels, boxes)]
@@ -498,7 +498,7 @@ class OwlViTImageProcessor(BaseImageProcessor):
             else:
                 img_h, img_w = target_sizes.unbind(1)
 
-            scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
+            scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
             boxes = boxes * scale_fct[:, None, :]
 
         results = []